Everyone agrees that it is better to tackle data quality issues and root causes rather than correct downstream data. When we talk about causes, we all have examples in mind because this is a problem that affects us all!
Possible origins of “bad quality”: some examples
- Input Errors: Input errors are a common source of data inaccuracy. Users sometimes make typing mistakes, enter data in the wrong field or select the wrong value from a list…Input errors can also be made voluntarily to save time or simplicity.
- Moving data out of applications: Employees can store data in local applications (e.g., Excel files) because they find it easier and faster or want to respond to an immediate need that the main tool does not respond to. In this case the data and version control is lost!
- Changes in the organization of services: they often lead to “accelerated” migrations of computer applications. These organizational changes lead in many cases to the appearance of duplicates and different “versions of the truth” for the same information.
- Data import and migration: When importing and migrating data must be done in a hurry, they will often create quality problems. Indeed, the data can be damaged or lost during the import if it is not conducted perfectly.
- Integration and exchange of data between applications: the interaction between the different applications of the information system is necessary to ensure the consistency of the data, but a minimal error during the transfers can easily cause problems on all the transmitted data.
- Shared Information between Services: For common information that should be unique, duplicates are often created because one service does not recognize the information created by another.
Clear rules and processes to avoid these sources of error
Implementing precise rules for document validation, dissemination, correction and analysis represents significant up-front costs. However, it is the only solution to avoid the direct costs of non-quality but also the process of bypassing and slowing down all processes. The poor quality of the data has a direct impact on the ability of the company to evolve by causing a strong inertia. Indeed, it is not only the errors that are critical but also their ability to spread and impact the entire company.